Introduction¶
MatchBench is a data integration ecosystem of libraries, which includes three main modules: Dataset interface, Integration algorithm, and Training/Evaluating module.
The library includes 20 data integration algorithms, covering four major categories:(Details).
Entity Resolution(ER) Entity Alignment(EA) Column Type Annotation(CTA) Schema Matching(SM)
We provide the support of 52 benchmark data integration datasets, 23 benchmarks for ER task, 20 benchmarks for EA task, 5 benchmarks for CTA task and 4 benchmarks for SM task (Details).
You can view MatchBench’s source code at MatchBench Repository.